p*****2
发帖数: 21240 | 1
monolithic
Fred
single
has
micro services can be host on same machine. communication is unnessesary to
be http |
|
z****e
发帖数: 54598 | 2 做到了soa之后，就没有必要纠结于框架的使用了
随便一个语言都可以实现soa
比如这1000行代码，用什么写都可以
只要最终包装成一个service
剩下的，自由选择
然后就是service自身的管理，会变成一个主要问题
其实这100个small service就需要做一个service manager一样的东西了
这个时候再通过层次来管理
殊途同归，这个东西在web service 1的时代有一个专门的东西，叫做uddi
hoho
monolithic
Fred |
|
g*****g
发帖数: 34805 | 3 1. It's a realistic question, not everything can be ran in parallel, and
when that's the case, async can only improve your system throughput, not how
quickly a single call can finish. Network latency is for real when you get
through hundreds of services. And for UI, 10ms and 100ms latency can be
easily told.
2. Single monolithic is worse. But you don't have to take it to another
extreme. In most cases, 1000 LOC is an API, not a service. 10K-50K is the
level I prefer, because a single engineer ca... 阅读全帖 |
|
n****1
发帖数: 1136 | 4 Google/Yahoo/Facebook? 普通非IT企业一般不会有这么大的人力物力和魄力做这么大
一个monolithic project吧, Hadoop架构?
看这位的形容，感觉最开始时的计划是很周密的，不用SOA肯定是代价太大。 俺估计他
们即使把整个项目用java重新写一遍, 还是不会用SOA。
check |
|
n****1
发帖数: 1136 | 5 China's railway business is still growing, maybe it will double or triple 10
years later. It is nice to see discussion on utilizing parallel system
instead of adding nodes blindly. But not able to scale out is a no-go,
period.
Your monolithic design make every single line of your code hardly reusable
in the parallel world. |
|
T********i
发帖数: 2416 | 6 我的铁路售票系统的架构已经说的很清楚了。
www.mitbbs.com/article_t/Programming/31298609.html
如果铁道部要求系统数据是紧耦合的，那么我的架构是性能最好的。所谓的分布式架构
不可能比我的性能好。没啥好讨论的。
事实上根据现在的情况看来。数据紧耦合确实是有这个要求的。至少能够购买联票带来
的用户体验是最好的。
这个系统抢票的核心确实是单机。提供最好的性能。而且杜绝超卖。多台单机串一串提
供跨DC的可靠性，应该是正常智商就能想到的。
其实，添加无限数量的state cache server预先过滤一下。只有state cache server显
示有票才放用户到核心去抢票。这个系统可以无限scale out的。这也是正常智商就能
想到的。但是，单机核心的重要性不因state cache的引入而有丝毫减弱。
这个系统设计和实现都很简单。不能因为你不会做，就假定别人做不出来。你的所谓
“Your monolithic design make every single line of your code hardly reusable
in the... 阅读全帖 |
|
n****1
发帖数: 1136 | 7 Your design is very similar to Apache Hbase, in which nodes are asymmetric.
Multiple master-node exist but only one master node is active, others are
ready to take over in case the active one fails.
I totally agree that CAP is satisfied if only eventual consistency is
required, because hbase already did that. Now please move on!
monolithic
granted
the |
|
z****e
发帖数: 54598 | 8 亮点在最后
http://www.infoworld.com/t/javascript/nodejs-keeps-stealing-rai
The replacement came about after Groupon decided it wanted to replace its
large monolithic application, which had trouble scaling, says Sean
McCullough, a software engineer at Groupon. The company already had the
JavaScript developer talent that could work with Node.js. "The goal was we
wanted to create new front ends just for each one of the sites, and Node was
well suited for the kind of engineers that we had and the perform... 阅读全帖 |
|
l**********n
发帖数: 8443 | 9 1. Treat new technology with healthy scepticism.
2. Be less trustful of corporate backing.
3. Prefer dedicated libraries to monolithic frameworks.
我对Angular2.0有很大期望。因为Angular1.0确实很牛逼。写代码的效率是其他
framework的十倍。 |
|
m*******k
发帖数: 178 | 10 我最讨厌的Eclipse，为什么他不骂？
Eclipse经常出现莫名奇妙的crash和hangs，即使不死机，也经常无法操作，也无错误
提示，多线程简直是笑话，所有操作都放进队列，编译任务经常无法取消，导致文件无
法保存，简直是垃圾中的极品。
因为和Torvalds没有竞争关系，他就不吭声了？和老印一个套路啊！！
如果说Eclipse不是一种技术，Emacs，GNOME和几个内核实现也未必是，最初版本的
linux，也就是在i386上面实现了几个基本的OS功能而已，在architecture设计方面也
未必有过人之处。Linux的monolithic design会让整个系统因为一个driver的小bug完
全瘫痪，内核安全性也没有保障，如果没有SELinux扩展，任何bug导致的进程拿到root
权限就可以完全攻破整个系统。很多攻击方式都是对linux最有效。他有这么多时间胡
说八道，如不多花一点时间研究一下如何把微内核的一些优势集成到Linux中去。 |
|
w***g
发帖数: 5958 | 11 1个100g的node，在性能上可能是好几倍10个10g的node的总和。
抛开分治需要写额外的代码不提(代码越多，出错的可能性也越多),
没有任何系统可以做到linear scale out。而且随着节点数量的增多，
mean time to failure会急剧缩短。所以总体开销其实会更大。
这个好处只有我自己知道，魏老师可能也知道，没法跟用AWS的人说明白。
我手头production system好几个，要是成天出错需要救火的话，
哪来那么多时间上网灌水。我有一个50多个节点的系统，实在是
数据量太大，没办法的那种。每台机器也有80～100g内存。隔
一两天就要处理一个错误。相比而言，那种两台机器做replicate的
monolithic系统，都是十天半个月都不需要碰一下。 |
|
g*****g
发帖数: 34805 | 12 I couldn't see how the Java to Scala shift made the difference for Gilt in
this article. It's all about Monolithic to Microservices, which you can do
in Java too. |
|
N********n
发帖数: 8363 | 13
我啥时说GENERIC是C#原创了？一开始我就说C#是在借鉴C++ TEMPLATE，跟JAVA有
茄子毛关系。JAVA那种MONOLITHIC RUNTIME搞法到C#上基本不适用。GLOBAL NAME
啥玩意风马牛不相及。 |
|
z****e
发帖数: 54598 | 14
那这个就是滥用了，一般继承关系不超过3层，一般超过3层就多半是屎坑
java再怎么严格，也不可能阻止滥写代码
代码不是越多越好，也不是越少越好
太少的话，人看不懂
太多的话，人也看不懂
一定要适度，凡事过犹不及啊
一般屎坑有个特点，就是文件特别长
超过1000行的class基本上都是维护的噩梦
老代码，尤其是monolithic时代的垃圾，5，6k行一个class很常见
现在都microservice了，所以一般都控制在200-500之间，很少超过500的
合理的文件长度在200-500之间
我自己控制在200左右，超过200我就开始分文件了
但是文件太多同样也有问题，一个方法就来一个文件
这种多半有问题
现在java的发展都已经开始减少继承关系了
abstract class我都不记得上次是什么时候用的了
一般都是一个interface，搭配一个impl class，搞定
甚至如果是同一台虚拟机里面的，直接就是class上
极少数说，有重合的地方，用上abstract class
而且现在都spring了，用interface感觉很怪异
绝大多数功能都是一个annotation就搞... 阅读全帖 |
|
b***i
发帖数: 3043 | 15 reserve(int start, int length, TicketPool *tp, Ticket *t)
里面这个t干啥的？
141 int startDelta = start - t->_start;
142 if (startDelta > 0) {
这个干啥？
我突然deja vu了，最近这个版有人提到吗，很多人争吵Micro Kernel， 还是
monolithic kernel好，整不出个所以然来，只有做出来看一看能不能用才行。 |
|
c***s
发帖数: 506 | 16 我知道啊。
很多人喜欢angular是因为two-way binding, 但这其实很多时候并不好用，如果不用这
个功能，那angular还有啥好处？
react一天就可以学会， 一个星其就可以高效率。angular要几天？
最最重要的是，front end 现在技术变得特别快，不要选一个monolithic的framework
，还是选library安全，转得快。 |
|
c***s
发帖数: 506 | 17 我知道啊。
很多人喜欢angular是因为two-way binding, 但这其实很多时候并不好用，如果不用这
个功能，那angular还有啥好处？
react一天就可以学会， 一个星其就可以高效率。angular要几天？
最最重要的是，front end 现在技术变得特别快，不要选一个monolithic的framework
，还是选library安全，转得快。 |
|
W***o
发帖数: 6519 | 18 上个周末花点时间把之前用node.js做的monolithic后台拆分成microservices，放在不
同的docker里面，发现这样维护起来更方便了，至少不会整个app瘫痪不work了 |
|
c******n
发帖数: 16666 | 19 应该说我漏了个先决条件——第二次refactor
单纯就快肯定是monolith最快
但是我想到如果要考虑到将来搞CI，可能还是在第一次refactor时候就改micro
service（第一次弄的时候肯定是能跑就行，真用mongo上schemaless感觉也算合理），
就如你说的去耦合，扩展起来也是独立进行的，
好处是不用全部推翻重来，然后对CI比较友好，缺点无非就是最后overhead比较高，那
就堆硬件呗。。直到有钱有人了再一个一个service来refactor |
|
d******c
发帖数: 2407 | 20 The Landscape of Parallel Computing Research - UC Berkeley EECS
1. Old CW: Power is free, but transistors are expensive.
· New CW is the “Power wall”: Power is expensive, but transistors are “
free”. That
is, we can put more transistors on a chip than we have the power to turn on.
2. Old CW: If you worry about power, the only concern is dynamic power.
· New CW: For desktops and servers, static power due to leakage can be 40%
of
total power. (See Section 4.1.)
3. Old CW: Monolithic uniprocessors ... 阅读全帖 |
|
m****x
发帖数: 2506 | 21 http://blog.devicerandom.org/2011/02/18/getting-a-life/
One of my first memories is myself, 5 years old, going to my mother and
declare to her, as serious as only children can be: “I will be a scientist.”
Yesterday night I was in my office in the Department of Chemistry at the
University of Cambridge packing my stuff, resolved to not go back to
research again -at least not in the shortcoming future.
What has gone wrong?
Not exactly my pathway (I finished Ph.D. quite well), but well, you get the
... 阅读全帖 |
|
a**********u
发帖数: 28450 | 22 【 以下文字转载自 WaterWorld 讨论区 】
发信人: monolith (on the road), 信区: WaterWorld
标 题: [请转生物版（第一天注册，只能发水版)问题急求] IP 问题，各位大牛请进
关键字: 生物版 IP问题
发信站: BBS 未名空间站 (Wed Apr 24 10:56:49 2013, 美东)
做一个Co-IP实验，检测骨架蛋白plectin(>500KD)与一个GTPase（A）的相互作用。内
源plectin,高表达A(A-flag)或 dominant negative A (DN-A-flag)。用plectin单抗做
IP，anti-flag blot 显示A-flag可以被co-IP下来，同样的co-IP GFP-negative
control 背景比较干净，而co-IP DN-A-flag有微弱信号（约为A-flag 的1/10）。
用做IP同样的antibody blot plectin,input 部分一切正常，显示相同的内源plectin;
问题出在 straight IP部分，同样的抗体同时blot同一块胶上... 阅读全帖 |
|
t******n
发帖数: 2939 | 23 【 以下文字转载自 WaterWorld 讨论区 】
发信人: monolith (on the road), 信区: WaterWorld
标 题: [请转生物版（第一天注册，只能发水版)问题急求] IP 问题，各位大牛请进
关键字: 生物版 IP问题
发信站: BBS 未名空间站 (Wed Apr 24 10:56:49 2013, 美东)
做一个Co-IP实验，检测骨架蛋白plectin(>500KD)与一个GTPase（A）的相互作用。内
源plectin,高表达A(A-flag)或 dominant negative A (DN-A-flag)。用plectin单抗做
IP，anti-flag blot 显示A-flag可以被co-IP下来，同样的co-IP GFP-negative
control 背景比较干净，而co-IP DN-A-flag有微弱信号（约为A-flag 的1/10）。
用做IP同样的antibody blot plectin,input 部分一切正常，显示相同的内源plectin;
问题出在 straight IP部分，同样的抗体同时blot同一块胶上... 阅读全帖 |
|
a*********n
发帖数: 602 | 24 The language of matrimony may often be invoked to describe corporate tie-ups
;
bust-ups too. But Rio Tinto's broken engagement with Chinalco of China
prompted
one of Beijing's more strident state-controlled newspapers to run an
editorial
last weekend that, in accusing the Anglo-Australian resources group of
infidelity,
made an elaborate contribution to the genre.
“Poor Chinalco prepared the wedding clothes but when the peach was ripe
somebody
else plucked it,” the Beijing Times opined. “Rio Tint... 阅读全帖 |
|
|
|
|
x****h
发帖数: 300 | 28 Bhattacharya, M.; Harold, M. P.; Balakotaiah, V. Mass-Transfer Coefficients
in
Washcoated Monoliths. AIChE J. 2004, 50, 2939-2955
Tsuchiya, T.; Ross, J. Application of Genetic Algorithm to Chemical Kinetics:
Systematic Determination of Reaction Mechanism and Rate Coefficients for a
Complex Reaction Network. J.Phys. Chem 2001, 105, 4052-4058.
Please send to x******[email protected]
Thank you very much. |
|
d********O
发帖数: 1196 | 29 Title:Monolithic, “Molded”, Porous Materials with High Flow
Characteristics for Separations, Catalysis, or Solid-Phase Chemistry:
Control of Porous Properties during Polymerization
1996, 8 (3), pp 744–750 DOI: 10.1021/cm950437j |
|
d********O
发帖数: 1196 | 30 Title: Kinetic Control of Pore Formation in Macroporous Polymers. Formation
of "Molded" Porous Materials with High Flow Characteristics for Separations
or Catalysis
1995, 7 (4), pp 707–715
DOI: 10.1021/cm00052a016
===============================================
Title; Preparation of Large-Diameter “Molded” Porous Polymer Monoliths and
the Control of Pore Structure Homogeneity
1997, 9 (8), pp 1898–1902
DOI: 10.1021/cm970204n
================================================
Please send to Drilling |
|
z*********4
发帖数: 146 | 31 Title; Preparation of Large-Diameter “Molded” Porous Polymer Monoliths and
the Control of Pore Structure Homogeneity
1997, 9 (8), pp 1898–1902
DOI: 10.1021/cm970204n
http://www.sendspace.com/file/mci4ox |
|
d********O
发帖数: 1196 | 32 Title: Preparation of macroporous methacrylate-based monoliths for
chromatographic applications by the Reactive Gelation Process
2010, Volume 1217, Pages 4675-4681
doi:10.1016/j.chroma.2010.04.077
Please send to D********[email protected] , if this is more convenient for you.
Thanks again! |
|
|
l***0
发帖数: 278 | 34 Preparation of combustion catalysts by wash coating alumina whiskers-
covered metal monoliths using a sol-gel method
Marcus F.M. Zwinkelsa, Sven G. Järsa and P. Govind Menona
Studies in Surface Science and Catalysis
Volume 91, 1995, Pages 85-94
link:
http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B8GYB-
B&_user=489835&_coverDate=12/31/1995&_rdoc=10&_fmt=high&_orig=browse&_or
ig
in=browse&_zone=rslt_list_item&_srch=doc-
info(%23toc%2342462%231995%23999089999%23661225%23FLP%23displa... 阅读全帖 |
|
x****h
发帖数: 300 | 35 Bhattacharya, M.; Harold, M. P.; Balakotaiah, V. Mass-Transfer Coefficients
in
Washcoated Monoliths. AIChE J. 2004, 50, 2939-2955
Tsuchiya, T.; Ross, J. Application of Genetic Algorithm to Chemical Kinetics:
Systematic Determination of Reaction Mechanism and Rate Coefficients for a
Complex Reaction Network. J.Phys. Chem 2001, 105, 4052-4058.
Please send to x******[email protected]
Thank you very much. |
|
x****h
发帖数: 300 | 36 1. Esposito, W. R. ; Floudas, C.A. global optimization of the Parameter
Estimation of Differential Algebraic systems, Ind. Eng. Chem. Res. 2000 39,
1291-1310
2.Young, L.C., The application of Orthogonal Collocation to Laminar Flow
Heat
and Mass Transfer in Monolith Converters,Ph.D. thesis, University of
Washington, Seattle (1974).
thanks a lot |
|
a*********r
发帖数: 14 | 37 对不起，我还真是给武大化学系留了面子。
莫非你就是其中一件丑事的主角，干嘛一触即发，破口大骂呢？
袁荃做的是还不错，但不意味武大化学系就有多好了。
举例来说，即使袁荃和另外一位新招的楚天特聘教授傅磊相比，也能看出武大化学系在其中有没有猫腻了。
首先补充一点：袁荃的老公叫廖蕾，也是武大物理系新招的楚天特聘教授，可以作为很好的对照体系。
比比三人的publication和经历，就能明显看出水平之比是:
廖蕾>袁荃>傅磊
但从袁荃却没有评上楚天特聘。
廖蕾
http://physics.whu.edu.cn/node/557
1. Y. Z. Long, L. Liao*, X. F. Duan, Z. Y. Fan*, Large-scale Integration of Semiconductor Nanowires for High Performance Flexible Electronics ACS Nano (In press)
2. L. Liao, J. W. Bai, R. Chen, H. L. Zh... 阅读全帖 |
|
a*********r
发帖数: 14 | 38 对不起，我还真是给武大化学系留了面子。
莫非你就是其中一件丑事的主角，干嘛一触即发，破口大骂呢？
袁荃做的是还不错，但不意味武大化学系就有多好了。
举例来说，即使袁荃和另外一位新招的楚天特聘教授傅磊相比，也能看出武大化学系在其中有没有猫腻了。
首先补充一点：袁荃的老公叫廖蕾，也是武大物理系新招的楚天特聘教授，可以作为很好的对照体系。
比比三人的publication和经历，就能明显看出水平之比是:
廖蕾>袁荃>傅磊
但从袁荃却没有评上楚天特聘。
廖蕾
http://physics.whu.edu.cn/node/557
1. Y. Z. Long, L. Liao*, X. F. Duan, Z. Y. Fan*, Large-scale Integration of Semiconductor Nanowires for High Performance Flexible Electronics ACS Nano (In press)
2. L. Liao, J. W. Bai, R. Chen, H. L. Zh... 阅读全帖 |
|
i*n
发帖数: 9793 | 39 J. Chamieh, C. Faye, V. Dugas, T. Moreau, O. Vandenabeele-Trambouze,C. Demes
may,
Preparation and full characterization of a micro-immunoaffinity monolithic c
olumn and its in line coupling with capillary zone electrophoresis with Ochr
atoxin A as model solute,
J.Chromatogr. A 1232 (2012) 93-100. |
|
j***h
发帖数: 4412 | 40 杭州9位海外高层次人才入选国家“千人计划”
2010年12月23日 04:10:23 浙江在线新闻网站
|人物名片|
聚光科技(杭州)有限公司董事长兼总工程师，浙江省环境与安全检测技术重点实验
室主任。获浙江大学光学仪器专业和美国斯坦福大学机械工程专业光电检测研究方向双
博士学位。
在国际半导体激光领域，王健的名字，在留学期间就贴上了“国际首创”的标牌。
在美国Stanford大学攻读博士学位期间，作为主要参与人,王健在国际上最先研制
成波分复用的cavity ringdown高灵敏度半导体激光光谱仪；最先研制开发出基于2.3微
米半导体激光器的CO遥测传感器；以第一作者在《测量科学与技术》发表的论文被该期
刊学术委员会评为2000年度唯一最优论文……
回国后，王健创办了聚光科技，组建了“光电测量技术研发中心”，逐步形成了一
支学历和专业结构配置合理、创新能力强的研发队伍，成为过程分析仪器领域国内规模
最大和国际上最优秀的创新团队之一。
王健带领聚光科技和杭州电子科技大学系统集成研究所两支创新团队，针对工业过
程和污染源在线监测等领域的检测需求，在国内最先研制成功“半导体激光在线气体分... 阅读全帖 |
|
s***l
发帖数: 2236 | 41 Monolithic Microwave Integrated Circuit
It's useful to make microwave circuit. Reduce the complaxity of
microwave strip calculation. |
|
s*d
发帖数: 309 | 42 monolithic microwave ic or just microwave ic?
they are very very different. |
|
y***q
发帖数: 98 | 43 The Applied Nano & Bio Photonics Group in Electrical Engineering Departm
ent of University of Arkansas has two Graduate Research Associate positi
on openings starting from either spring 2009 or fall 2009. The research
projects include: i) design/fabricate/characterize/package advanced sem
iconductor optoelectronic devices for biomedical imaging applications; i
i) develop novel optoelectronic devices using the innovative material in
tegration techniques such as monolithic II/VI, IV (SiGeSn) and |
|
g*******r
发帖数: 140 | 44 C.F. Lee and P.K.T. Mok, “A Monolithic Current-Mode CMOS DC-DC Converter
with On-Chip Current-Sensing Technique,” IEEE Journal of Solid-State
Circuits, Vol. 39, No. 1, pp. 3-14, Jan 2004.
I think this is a good paper to start with. |
|
L****a
发帖数: 572 | 45 give you some small names on Analog/Power/mixed signal :
Advanced Analogic,
Microsemi Corporation
ON Semiconductor
Power Integrations
Silicon Laboratories
Monolithic Power Systems
however, most of the analog company are not in a good shape
at this time. |
|
g*****g
发帖数: 3623 | 46 ft
基本上一网打尽了。。。
power one不是 IC company，和 Cherokee International 一样早就一直在亏
t
不知道 emerson network power怎么样
不知道 volterra，Monolithic Power Systems， power integration，silicon lab，
这些小一点的怎么样
year) |
|
L****a
发帖数: 572 | 47 你说的是 LED backlight 吗? 否则光LED 也太广泛了.
你说的资料也太广泛了, 你说的是面板的工艺, 还是驱动电路, 还是激光器件.
我知道国内有些产商有生产面板, 但是都是不死不活的. 驱动电路, 主要是
analog 和 power IC. 美国很多 analog IC 产商都做, 比如 monolithic power
system, Catalyst, Advanced Analogic... 国内如果开始做的话, 也就是在价格
上取得优势而已.
查。 |
|
y***q
发帖数: 98 | 48 The Applied Nano & Bio Photonics Group in Electrical Engineering Departm
ent of University of Arkansas has two Graduate Research Associate positi
on openings starting from either spring 2009 or fall 2009. The research
projects include: i) design/fabricate/characterize/package advanced sem
iconductor optoelectronic devices for biomedical imaging applications; i
i) develop novel optoelectronic devices using the innovative material in
tegration techniques such as monolithic II/VI, IV (SiGeSn) and |
|
c*******y
发帖数: 3008 | 49 我们主要从事Monolithic Refractory Nanotechnology 的manufacturing company
based in chicago, we have 3 plants in the U.S, 1 in UK, 1 in Tianjin.
our product lines：
•Abrasion Resistant Products
•Aluminum Industry Products
•Blast Furnace Reprofile Products
•Casthouse Products
•Copper Industry Products
•Delta Products
•Ferro Alloy Products
•Other Applications
•Glass Industry Products
•Reheat Furnace Pumpables
•Reheat Furnace Shotcretes
our pr... 阅读全帖 |
|
m***n
发帖数: 266 | 50 来自主题: Macromolecules版 - 紧急求助 drug delivery
monolithic system. dissolved active drug in matrix
desorption of drug can be expressed as early time and late time.
desorption from a slab with thickness of l can be expressed by
two series, which is derived in Crank's book "mathematics of diffusion".
现在找不到这本书，下周一需要交推导过程。哪位知道如何从建立模型到推出这
两个series的表达式。
大致为： Mt/Mo=1- infinite
__
\ 8exp[-D(2n+1)super2paisuper2t/l2}/(2n+1)2pai2
/__
没法表示清楚这个公式。大致意思吧。多谢多谢！ |
|
